[Tcl-bugs] [ tcl-Bugs-474335 ] logic of matherr is backwards

[<no...@so...>]

Bugs item #474335, was opened at 2001-10-23 20:45
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=110894&aid=474335&group_id=10894

Category: 48. Portability Support
Group: 8.4a4
Status: Open
Resolution: None
Priority: 5
Submitted By: Don Porter (dgp)
Assigned to: Mo DeJong (mdejong)
Summary: logic of matherr is backwards

Initial Comment:

A simple look at the cvs history of 
unix/tclMtherr.c shows the introduction of
a bug:

cvs diff -r core-8-0-5 -r core-8-1-0 tclMtherr.c
...
matherr(xPtr)
     struct exception *xPtr;    /* Describes error that
occurred. */
 {
-    if (!tcl_MathInProgress) {
+    if (TclMathInProgress()) {
        return 0;
...

The "!" did not survive the transition from
access of a global variable to call of an internal
function.  The logic of matherr() became reversed.
The obvious fix is to restore the proper logic.

It's worth pausing though and asking why this
bug hasn't been noticed in the 3 years since it
was introduced.  I am unable to come up with
any script on any platform that exposes this error.

It seems that Tcl behaves the same whether the
logic is forwards or backwards.
It would seem then that Tcl would also behave the
same if the logic were completely removed.  I welcome
second opinions on this, but my inclination would
be to remove the matherr() routine altogether
unless someone can demonstrate it's serving a
useful purpose somewhere.  (And how useful can
it be if it's been wrong since Tcl 8.1.0 ?)


----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=110894&aid=474335&group_id=10894